J. Paleont., 75(1), 2001, pp. 34±45 Copyright ᭧ 2001, The Paleontological Society 0022-3360/01/0075-34$03.00

TRIASSIC ROOTS OF THE AMPHIASTRAEID SCLERACTINIAN CORALS

EWA RONIEWICZ AND JAROSèAW STOLARSKI Instytut Paleobiologii, Polska Akademia Nauk, Twarda 51/55, 00-818 Warszawa, Poland, Ͻ[email protected]Ͼ, Ͻ[email protected]Ͼ

ABSTRACTÐThe Early Carnian (Upper Triassic) phaceloid coral originally described by Volz (1896) as Hexastraea fritschi, type species of Quenstedtiphyllia Melnikova, 1975, reproduced asexually by ``Taschenknospung'' (pocket-budding), a process documented herein for the ®rst time. This type of budding is recognized only in the Amphiastraeidae, a family thus far recorded only from Jurassic- Cretaceous strata. Similar to amphiastraeids, Quenstedtiphyllia fritschi (Volz, 1896) has separate septal calci®cation centers and a mid- septal zone built of serially arranged trabeculae. The most important discriminating characters of the new amphiastraeid subfamily Quenstedtiphylliinae are one-zonalendotheca and radial symmetry of the corallite in the adult stage (in contrast to two-zonal and bilateral symmetry in the adult stage in Amphiastraeinae). Quenstedtiphyllia fritschi shares several primitive skeletal characters (plesiomorphies) with representatives of Triassic Zardinophyllidae and, possibly, Paleozoic plerophylline rugosans: e.g., thick epithecal wall and strongly bilateral early blastogenetic stages with the earliest corallite having one axial initial septum. To interpret the phylogenetic status of amphiastraeid corals, we performed two analyses using plerophylline rugosans and the solitary scleractinian Protoheterastraea, respec- tively, as the outgroups. The resulting phylogenetic hypotheses support grouping the Zardinophyllidae with the Amphiastraeidae in the clade Pachythecaliina (synapomorphy: presence of pachytheca). Taschenknospung is considered an autapomorphy for the Amphiastraei- dae. This study is the ®rst attempt to analyze the relationships of the Triassic corals cladistically.

INTRODUCTION Subsequent stages of the earliest phase of budding were traced at intervals of 2 mm due to a lack of material. CLERACTINIAN CORALS arose during the Middle/Late Triassic  S as four major microstructural groups (Roniewicz and Mo- Institutional abbreviations. MGUWr, Geological Museum, rycowa, 1993; see also Fig. 1): 1) mini-trabecular (e.g., reimani- Wrocøaw University, Wrocøaw, Poland; ZPAL, Institute of Paleo- phylliids); 2) thick-trabecular (e.g., actinastreids, pamiroseriids); biology, Warsaw, Poland; GMH, Geiseltal Museum, Halle, Ger- 3) fascicular (ϭnon-trabecular, i.e., stylophyllids and gigantostyl- many; UMIP, University of Montana, Institute of Paleontology, ids); and 4) pachythecal (zardinophyllids). Phylogenetic relation- USA. ships among these groups are poorly understood, but differences MORPHOLOGY OF THE SKELETON in microstructural characters, especially among pachythecal, sty- Zardinophyllids are solitary, i.e., Pachythecalis Cuif, 1975; lophyllid, and mini/thick-trabecular groups, suggests a polyphy- Zardinophyllum Montanaro-Gallitelli, 1975; and phaceloid, i.e., letic origin (Roniewicz and Morycowa, 1993). A polyphyletic or- Pachydendron Cuif, 1975; Pachysolenia Cuif, 1975 (probably the igin of Scleractinia is also suggested by phylogenetic analyses of senior subjective synonym of Lubowastraea Melnikova, 1986, a DNA sequences of shallow-water scleractinians (Romano and form with corallites apposed to each other). A very thick wall is Palumbi, 1996; Veron et al., 1996). a predominant skeletal character. Septa are non-exsert and usually Particularly intriguing is the origin and evolution of Triassic located deep in the calice. Adult coralla may have quasi-radial corals that have skeletal architecture strikingly similar to that of symmetry, whereas initial and juvenile stages are often strongly the late Paleozoic rugosans. One such coral group is the Zardi- bilaterally symmetrical. In solitary Pachythecalis and Zardino- nophyllidae, which resembles plerophylline rugosans by having phyllum, intracalicular space is ®lled during the ontogeny with an identical initial corallite ontogeny, an extremely thick corallite compact sclerenchyme (stereome), whereas in phaceloid Pachy- wall (in relation to the calice diameter), rhopaloid septa, and many solenia and Pachydendron this space is ®lled with tabuloid dis- other aspects of corallum morphology and microstructure (see sepiments. Cuif, 1975b; Cuif and Stolarski, 1999; Stolarski 1999). Zardino- Protoheterastraeids are restricted herein to solitary Protohet- phyllids are represented by four fossil genera: Pachythecalis Cuif, erastraea Wells, 1937, represented by the lectotype of the type 1975 (early Norian), Pachysolenia Cuif, 1975 (early Carnian± species Protoheterastraea leonhardi (Volz, 1896). Protoheteras- Norian), Pachydendron Cuif, 1975 (early Norian±Rhaetian), and traea has a relatively thick epithecal wall, tabuloid dissepiments, Zardinophyllum Montanaro-Gallitelli, 1975 (early Carnian± and radial symmetry. Cerioheterastrea, another genus assigned Rhaetian). Analysis of the skeletal structure and budding of the by Cuif (1976) to the ``Protoheterastraea-group'', differs signif- amphiastraeid Quenstedtiphyllia fritschi (Volz, 1896) has yielded icantly from Protoheterastraea in septal structure and endotheca, new data, which have encouraged us to renew discussion of the and in our opinion, these two genera should not be grouped to- origin and evolution of Zardinophyllidae. gether. Ladinian solitary/phaceloid corals with thick epitheca and tabular endotheca that resemble Protoheterastraea have been il- MATERIAL AND METHODS lustrated by Deng and Kong (1984, pl. 3, ®gs. 7, 9; identi®ed as Elasmophyllia and, respectively, Pinacophyllum). We examined the holotype of Quenstedtiphyllia fritschi (Volz, Amphiastraeids are solitary (Cheilosmilia Koby, 1888), pha- 1896), the only known specimen of this species. The specimen celoid (Amphiaulastrea Geyer, 1955; Aulastraea Ogilvie, 1897; (colony fragment, size 80 ϫ 70 ϫ 40 mm) was collected in the Mitrodendron Quenstedt, 1881; Pleurophyllia de Fromentel, lower Carnian St. Cassian Beds in the Dolomites (Southern Alps), 1856; Hexapetalum EliasÏova, 1975; Pseudopisthophyllum Geyer, and is housed at the Geiseltal Museum in Halle (Germany). Its 1955; Hykeliphyllum EliasÏova, 1975); cerioid (Pleurostylina de original aragonite mineralogy is preserved, thus enabling micro- Fromentel, 1856); or pseudocerioid with separated polygonal cor- structural and ontogenetic (blastogenetic) observations. We base allites (Amphiastraea Etallon, 1859). The corallites are usually our observations on thin sections (12 slides), acetate peels taken strongly bilaterally symmetrical. The wall is thick. The endotheca from transverse and longitudinal sections of the colony, and a is built of tabuloid dissepiments, which peripherally may pass into series of peels made every 0.1 mm from the budding corallite. a zone of vesicular dissepiments (Melnikova and Roniewicz, 34 RONIEWICZ AND STOLARSKIÐTRIASSIC AMPHIASTRAEID CORALS 35

FIGURE 1ÐFour basic skeletal microstructures of the Triassic Scleractinia. 1, Minitrabecular (e.g., Protoheterastraea). 2, thick-trabecular (e.g., Chondrocoenia) 3, fascicular (e.g., Stylophyllum). 4, pachythecal (e.g., Pachydendron). After Roniewicz (1996).

1976). For the present purposes, we have not included in this fascicles of ®bers with various dimensions to well delimited mod- analysis a group of Jurassic scleractinians traditionally classi®ed ules without axes (Fig. 2.1, 2.3). Generally, fascicles that form with amphiastraeids in the suborder Amphiastraeina Alloiteau, wall modules extend into the septa. 1952: i.e., Donacosmiliidae, Carolastraeidae, Intersmiliidae. They Septa.In zardinophyllids, calci®cation centers of the mid-sep- are known only from completely recrystallized specimens, and tal zone are, as a rule, not separated. They form a homogeneous are generally very poorly characterized, such that many of their zone composed of microcrystalline material in longitudinal and characters would be represented in a phylogenetic analysis as transverse sections of septa (vertical and horizontal sections are missing values, producing unreliable results. required to detect possibly horizontally, or respectively, vertically Quenstedtiphyllia fritschi (Volz) has a phaceloid growth form. arranged septal trabeculae). In transverse section, ®bers thicken- The corallites are radially symmetrical with penta- and hexameral ing the septum are often oriented strongly oblique to the mid- septal con®gurations composed of regularly distributed septa in 4 septal zone, indicating predominant axial accretion of the septum orders. The most striking morphological character is a thick wall (e.g., Pachythecalis major, Pachydendron microthallos, see Cuif, (Fig. 2). The calicular lumen is ®lled by tabuloid (one-zonal) en- 1975b, ®gs. 4, 6b respectively). Septal accretion was often rhyth- dotheca. mic, and its successive phases can be distinguished in transverse section as superimposed secretional units (e.g., in Pachythecalis MICROSTRUCTURE OF THE SKELETON Cuif, 1975b, ®g. 3b, d). To date, longitudinal sections of Pachy- Wall.Zardinophyllidae and Amphiastraeidae have a unique solenia cylindrica (type species of the genus) have not furnished type of epithecal wallÐa pachythecaÐwith an internal layer (i.e., us with microstructural details of septa. Thus, it is unclear whether epithecal, or rather pachythecal stereome) built of radially orient- this species has non-separated calci®cation centers of the mid- ed equal-sized fascicles of ®bers (``modules'' of Roniewicz and septal zone. In transverse section, the septa of P. cylindrica seem Stolarski, 1999). The calci®cation centers of these fascicles are at to be composed of several axially directed secretional units that the circumference of the corallite (see Cuif, 1975b, ®g. 9b), or have non-separated calci®cation centers (Cuif, 1975b, ®g. 9b). form the axes of the fascicles (in this case, ``modules'' form hor- However, a similar observation can also be made in transverse izontal trabeculae; see Cuif, 1975b, ®g. 5b; herein Fig. 2.4). In sections of septa with the trabeculae directed axially; such axi- transverse sections of various zardinophyllids, ®brous structures alward, serially arranged trabeculae have been actually described of the wall and septa are usually distinctly separated by a suture in Carnian Pachysolenia primorica (see Iljina, 1984, ®g. 32.1b). (indicating a temporal/spatial gap in ontogeny of these struc- In the mid-septal zone of Quenstedtiphyllia fritschi, calci®ca- turesÐe.g., Pachythecalis, Pachydendron, and PachysoleniaÐ tion centers are separated, and the septa are built of serially ar- Cuif, 1975b, ®gs. 3d, 4, 6b, 9b). Rarely, the ®bers of the epithecal ranged trabeculae that are oriented centripetally (Figs. 2.2, 3.1). stereome continue into radial elements, especially those of higher The trabeculae, ca. 40 ␮m in diameter, are visible in longitudinal orders (indicating no temporal/spatial gap in the ontogeny of these and oblique sections. Their tips form a sharp denticulation on the structures). internal septal edge. Traces of axially directed trabeculae forming In Protoheterastraea, the wall is relatively thick, and built of a regular denticulation on the axial edge of the septum were also radially oriented ®bers. The fascicles of ®bers are arranged into discerned in Q. mardjanaica Melnikova, 1975. large units which are, however, much larger and less delimited In Protoheterastraea, calci®cation centers of the mid-septal than typical zardinophyllid ``modules.'' Calci®cation centers are zone are well separated (ca. 30±45 ␮m), and septal trabeculae located exclusively on the perimeter of the wall. An identical type are arranged in a fan system. of thick epithecal wall (mature epitheca) was recognized in Re- Budding.The most common modes of asexual increase of cent Gardineria (Stolarski, 1996a, ®g. 5G). solitary and phaceloid epithecate scleractinians are: 1) corallite In Quenstedtiphyllia fritschi, the organization of ®bers of the division, 2) lateral budding, and, 3) ``Taschenknospung'' (ϭpock- epithecal stereome shows considerable variation, ranging from et budding). In corallite division (Fig. 5.1), the parental corallite 36 JOURNAL OF PALEONTOLOGY, V. 75, NO. 1, 2001 splits into more or less equal daughter corallites. In the splitting into pachythecal wall. ``Modular'' style of biomineralization (with area, opposite septa of the parental corallite either join and form horizontal modules having circumferential calci®cation centers) is an incipient wall that separates two daughter corallites, or are not thus changed into trabecular one (with vertical trabeculae having involved in corallite division. In lateral budding (Fig. 5.2), new axially arranged calci®cation centers). The ``modular'' biominer- buds grow centrifugally from the parental corallite almost from alization style is restored during the subsequent growth after the the beginning of the corallite division, whereas in Taschenknos- separation of daughter and adult corallites (in the blastogeny of pung (Fig. 5.3) new buds enlarge at the expense of calicular space corals with a trabecular wall, e.g., Madrepora, trabecular style in the parental corallite until they ultimately emerge. Long-lasting calci®cation of the wall is uninterrupted; e.g., Stolarski, 1996b, parallel growth of the parental and daughter corallites results in pl. 6.3). In the skeletal ontogeny of extant corals we found the an increase in the diameter of the latter (Fig. 6). following examples of analogous changes in biomineralizational The term Taschenknospung was used by Ogilvie (1896) to de- styles: differentiation of separated septal calci®cation centers on scribe a mode of budding in the Jurassic (Tithonian) amphias- the basal plate, the primary layer of which consists of non-sepa- treids, and by Volz (1896) for the Triassic Hexastraea fritschi. rated calci®cation centers (e.g., in Porites lutea, see Jell, 1981), The ®rst comprehensive description of Taschenknospung was or formation of the epithecal wall (non-separated calci®cation based on serial sections of the Late Jurassic amphiastreids Pleu- centers) in rejuvenated corallites that otherwise have trabecular rophyllia and Mitrodendron (see Roniewicz, 1966; Melnikova walls (e.g., in Polycyathus muellerae, see Roniewicz and Stolar- and Roniewicz, 1976). Typical Taschenknospung was described ski, 1999, ®g. 3B). Similarly, in the ontogeny of azooxanthellate only for the Amphiastraeidae. It is also possible that the Hetero- Schizocyathus and Pourtalocyathus, the wall may change from coeniidae budded by Taschenknospung; however, Koøodziej's typical smooth epitheca (with non-separated calci®cation centers) (1995) interpretations are not supported by serial sections, and to hispidotheca (with separated calci®cation centers that are found young ``buds'' may represent rejuvenated corallites. in spherulite-like bodies; Stolarski, 2000). In phaceloid Zardinophyllidae, new buds are formed by lateral Changes in biomineralization styles that occur in skeletal ontogeny budding with new corallites growing centrifugally (Pachydendron provide important clues about possible constraints on microstructural microthallos Cuif, 1975, pl. 14, ®g. 2; Pachysolenia cylindrica evolution of the scleractinian skeleton. They suggest that borders Cuif, 1975, ®g. 8a). between microstructural groups may not be so strict as suggested In Q. fritschi, we af®rmed typical Taschenknospung budding, before (Roniewicz and Morycowa, 1993), and that phylogenetic re- with new buds initially growing axialward, at the expense of the lationships are possible between some of these groups, e.g., pachy- space of the parental calice, after which growth becomes centrif- thecaliines and protoheterastraeids as postulated herein. Nevertheless, ugal (Figs. 4.3, 4.4 5.1±5.4). One to three new calices may orig- the disparity of the Middle Triassic scleractinians is too large for us inate simultaneously in the pachythecal stereome of the parental to assume their monophyletic origin, although it is possible that the calice. We distinguished the following stages of Taschenknospung number of ancestral groups was smaller than the eight groups sug- budding in Q. fritschi (Fig. 5): i) appearance of a ``disturbance'' gested by Roniewicz and Morycowa, 1993, or the ten (or more?) in the radial arrangement of fascicles of ®bers of the epithecal groups suggested by Veron, (1995). stereome; ii) appearance of irregularly arranged vertical trabecu- lae that form a semilunar palisade of the wall delimiting the future ADAPTIVE SIGNIFICANCE OF TASCHENKNOSPUNG calice; iii) formation of a small, elongated cupule (``pocket'') ex- Phaceloid corals with an epithecal wall and endotheca form a ternally to semilunar wall palisade, which is divided into two separate category of ``pseudocolonial'' corallum organization, lying compartments by a relatively thick, primary ``axial'' septum; iv) between solitaries and colonies (see Coates and Jackson, 1985; Ro- transformation of the primary septum into a free and thin septum; sen, 1986; Roniewicz and Stolarski, 1999). In solitary corals, the and v) development of one or two septa to the right and left of coelenteron forms a sort of mesenterially folded sack with one major the primary septum, and the nearly simultaneous appearance of ori®ce (monostomatous condition). If solitary corals reproduce asex- incipient septa on the other side of the cupule. ually by transverse or longitudinal division, then the coelenteral space The septa of the daughter calice are independent from those of is shared among parental and daughter coralla for the duration of the parental one. Observation of several sections testi®es that the this process. Just after division, skeleton and soft tissue become, as buds are growing within the limits of the parental calice for some a rule, permanently separated (e.g., Yamashiro and Yamazato, 1987). time, and that they emerge at the stage with a highly advanced In colonial corals, the coelenteron forms a complicated network that septal apparatus. Melnikova (1975) reported Taschenknospung in connects better or less individualized polyps (polystomatous condi- Quenstedtiphyllia mardjanaica, and later (Melnikova, 1986, ®g. tion or typical colonies with well delimited polyps and corallites). In 7:1b) ®gured a large, daughter corallite with numerous septa that colonies, the soft tissue and the skeleton continue between corallites. already emerged from the parental corallite. In phaceloid corals, coelenteral space is shared between parental and The same stages of Taschenknospung as described above in Q. daugther polyps only early in blastogeny. Formation of the epithecal fritschi were distinguished in Amphiastraeidae: in Mitrodendron wall on the entire circumference of the daughter corallite and ap- (see Roniewicz, 1966; herein Fig. 6), and Pleurophyllia (see Mel- pearance of the ®rst tabuloid dissepiments suggest that the soft-tissue nikova and Roniewicz, 1976, ®gs. 1, 2). connection between daughter and parental polyps is cut off but the Budding was not observed in Protoheterastraea leonhardi. A skeleton remains connected. Separation of the young polyp from the branching coral assigned by Cuif (1973) to P. leonhardi and in- parental one was probably relatively fast in phaceloid forms with creasing by subequal septal division (Cuif, 1973, ®gs. 26, 27), lateral budding, whereas in those with Taschenknospung, it appears actually represents a different taxon. to be prolonged. In Q. fritschi, daughter corallites are separated from the parental one by the trabecular wall (Figs. 3, 4). This kind of wall BIOMINERALIZATION AND ITS IMPLICATIONS is formed by ectoderm that continues between polyps. Despite the During early blastogeny of the corallum of Quenstedtiphyllia proximity and continuation of certain parts of the soft-tissue, daugh- fritschi, two important microstructural changes occur: a) fascicles ter polyps do not inherit the mesenterial system from the parental of ®bers that form the inner pachythecal layer differentiate into polyp as suggested by the highly underdeveloped septal system of vertical trabeculae of the primordial wall that separates parental the daughter coralla. Because development of mesenterial cycles cor- and daughter corallites (Fig. 3.1), b) during the subsequent divi- relates with the formation of the tentacular crown, young polyps sion of the corallites, the typical trabecular wall transforms again most likely had only few tentacles, and were defended primarily by RONIEWICZ AND STOLARSKIÐTRIASSIC AMPHIASTRAEID CORALS 37

FIGURE 2Ð1±3, Quenstedtiphyllia fritschi (Volz, 1896). Lower Carnian, St. Cassian Formation, Dolomites, Italy. GMH, holotype; 1, transverse section showing a pachythecal wall with well-delimited modules, and pentameral symmetry of the septal apparatus; 2, longitudinal section of a septum with obliquely cut trabeculae; 3, transverse section of the pachythecal wall; the modules have calci®cation centers situated on the wall circumference (arrows). 4, Pachysolenia cylindrica (Cuif, 1975). Lower Norian, Alakir Cay, Taurus Mountains, Turkey. UMIP 196100. Transverse section of the pachythecal wall formed by horizontal modules with calci®cation centers in their axis (i.e., trabecular organization); see arrows.

tentacular crown of their parental polyp. Because tentacles are used but also vulnerable to predators. Some analogy to these two types in feeding, we suggest also that a young, underdeveloped polyp was of asexual reproduction are brooding vs. broadcast spawners strate- dependent for a long time on the parental polyp, not only by the gies recognized among generations reproducing sexually (e.g., Rink- protection afforded by its tentacular crown but also, to some extent, evich and Loya, 1979). by transfer of nutrients through a shared coelenteron. Ultimate sep- aration from the parental polyp, which is indicated by the circle of PHYLOGENETIC ANALYSIS epithecal wall, took place at the point of development of very well developed digestion (mesenteria) and defense (tentacular crown) sys- Selection of outgroups.Appearance of only one axial septum tems. We contrast this budding strategy with typical lateral budding, in the early blastogenetic stage of Quenstedtiphyllia followed by in which shortly after the division young polyps become independent the strongly bilateral (two-by-two) insertion of the next septa, as 38 JOURNAL OF PALEONTOLOGY, V. 75, NO. 1, 2001

FIGURE 3ÐQuenstedtiphyllia fritschi (Volz, 1896). Lower Carnian, St. Cassian Formation, Dolomites, Italy. GMH, holotype. 1, Longitudinal thin section of a septum showing tips of trabeculae on its internal border; 2, transverse thin section of the wall that is developed between the parental and descendant corallite during Taschenknospung. Calci®cation centers of the zigzag mid-wall zone are slightly diagenetically altered (microcrys- talline material is no longer recognized), thus it is not clear whether all centers were well separated (as still recognizable in some parts of this section) or whether well- and non-separated calci®cation centers coexisted. 3, 4, peels of budding corallites in transverse section, showing young corallites originating by Taschenknospung. The walls dividing the corallites are formed by thick, vertical trabeculae. well as a thick epithecal wall suggest close phylogenetic relation- transitional forms between plerophylline and pachythecaliine cor- ships among Quenstedtiphyllia, amphiastraeids, zardinophyllids, als (see Oliver, 1981; Fedorowski, 1997). However, in the ontog- and, possibly, plerophylline rugosans. To test this hypothesis (see eny of Zardinophyllum, phases of septal insertion occur that are also Melnikova and Roniewicz, 1976; Stolarski, 1996a), we have strikingly similar to those expected from such an ``ideal transi- ®rst performed phylogenetic analysis using the discussed taxa and tional form'' (Stolarski, 1999; see also discussion below con- choosing plerophylline rugosans as the presumed outgroup. The cerned characters III to V), and we consider this as a strong ar- concept of using plerophylline rugosans as an outgroup is still gument supporting the hypothesis of the direct plerophylline de- considered controversial, because no early Triassic skeletonized scent of the pachythecaliine corals. We argue also that disparity anthozoans have been recorded that could show an ideal series of between rugosans and scleractinians in skeletal mineralogy may RONIEWICZ AND STOLARSKIÐTRIASSIC AMPHIASTRAEID CORALS 39

FIGURE 4ÐQuenstedtiphyllia fritschi (Volz, 1896). Lower Carnian, St. Cassian Formation, Dolomites, Italy. GMH, holotype. Early stages of Tas- chenknospung budding documented by transverse serial sections preserved as peels. The process is herein traced over a distance of 1.1 mm. 1, Initial stage of formation of the wall between the bud and parental calice. The ®rst vertical trabeculae of the new wall, formed in the widened pachytheca of the adult form, are circled (and marked with arrows); 2, early stage of calice formation, with thick axial septum dividing the calicular lumen, and with a smooth internal wall surface. 3, shortening of the axial septum and its transformation into a septal blade. Other septa appear symmetrically on both sides of the axial septum. The internal surface of the wall on the opposite side is covered with low ridges marking the position of appearing septa; 4, development of the septal apparatus. The axial septum splits into two septa, and some septa appear on the opposite side. The diameter of the bud signi®cantly increases. Measurements (h) give vertical height above ®rst peel. All scale bars are 250 ␮m. be less important than is traditionally suggested (Oliver, 1981); an outgroup of the pachythecaliines. Solitary/phaceloid corals see discussion on character I. with thick epitheca and tabular endotheca that resemble Proto- In the second analysis, we have excluded plerophylline rugo- heterastraea (see notes above in the chapter: Morphology of the sans from the outgroup based on the traditional assumption that Skeleton) are rare but distinct components among earliest scler- pachythecaliines are convergent with them (i.e., Oliver, 1981, p. actinian faunas, consisting mostly of colonial corals with rela- 399 considered Zardinophyllum ``an aberrant scleractinian''). This tively well integrated colonies. Protoheterastraea is thus used as reasoning is supported by the lack of unquestionable pachythe- an outgroup in the second analysis. caliines in the earliest Middle Triassic coral assemblages (the Lad- By choosing different outgroups, the character polarization is inian Zardinophyllum illustrated by Deng and Kong, 1984, pl. 3, almost completely reversed in two analyses (compare ancestral ®g. 1 represents, most likely, an isolated tube of the lemniscater- conditions on Fig. 7). ine hydrozoan Cassianastraea Volz, 1896), although this may Characters.We used the following 10 macro- and microstruc- well be interpreted as sampling bias (even Carnian±Norian zar- tural characters (Table 1). The characters of plerophylline rugo- dinophyllids constitute only a subordinate group in coral assem- sans are those used by Hill (1981) for description of Plerophyllina blages and thus could be overlooked in generally poorly preserved Sokolov, 1960; microstructural and ontogenetic data are extracted and rare Anisian±Ladinian coral faunas). By excluding plero- mainly from Schindewolf (1942) and Iljina (1984). phyllines, few genera remain that can be reasonably considered Character I.Skeletal mineralogy. Coralla of all scleractinians 40 JOURNAL OF PALEONTOLOGY, V. 75, NO. 1, 2001

FIGURE 5ÐCommon budding strategies in phaceloid epithecate corals (®gured models do not represent any particular taxon). 1, Taschenknospung: new bud enlarges at the expense of calicular space in the parental corallite until ultimately emerges; 2, lateral budding: new bud grows centrifugally from the parental corallite almost from the beginning of the corallite division; 3, corallite division: parental corallite splits into equivalent daughter corallites. thus far analyzed are entirely aragonitic (e.g., Cairns, 1995; Sto- This character is removed from the analysis using Protoheter- larski, 1996a; Cuif and Dauphin, 1998). Traces of calcite recorded astraea as an outgroup, since aragonitic mineralogy is autapo- from calci®cation centers (Constantz and Meike, 1990) seems to morphic for Scleractinia. be of diagenetic origin (see Cuif and Dauphin, 1998; Sorauf and Character II.Budding. The solitary state (absence of the bud- Cuif, 1999). In contrast, skeletons of the majority of rugosans ding) is found in many plerophyllines. However, some of them (including Plerophyllina) are calcitic [NoteÐWendt (1990) sug- form phaceloid colonies by lateral budding (e.g., Calophyllum gested that in addition to Numidiaphyllum, redescribed as a Perm- gemmatum Iljina, 1984, ®g. 58). We assumed the polymorphic ian scleractinian by Ezaki (1997), the skeleton of the typical ru- character state (solitary/lateral budding) to be ancestral. gosan Ipciphyllum arnouldi was also originally aragonitic, but this The solitary character state is found in Protoheterastraea, and information needs veri®cation]. Global geochemical ¯uctuations, in the second analysis, the solitary character state is used as an- especially changes in Mg/Ca ratio of seawater, have been sug- cestral. gested as guiding the evolution of skeletal mineralogy in many Character III.One/two or axial initial septa. In plerophylline groups of organisms (e.g., Railsback and Anderson, 1987; Stanley rugosans, the initial ontogeny is fully comparable, including ap- and Hardie, 1998). One may suggest that the radical shift in Mg/ pearance of a one/axial septum as the ®rst radial element of skel- Ca ratio of seawater across the P/T boundary, which favored evo- etal ontogeny/blastogeny (compare Iljina, 1984; Stolarski, 1999). lution of aragonite secretors in the Triassic, also played an im- Presence of one/two or axial initial septa is thus taken as the portant role in the skeletal mineralogical transformation between plesiomorphic state in the ®rst analysis. We assume that the early the plerophyllines and the pachythecaliines. In our opinion how- ontogenetic stages of zardinophyllid and the early blastogenetic ever, geochemical factors themselves were not a main triggering stages of amphiastraeids are similar. In the earliest blastogenetic mechanism of the suggested mineralogical change in biocalci®- stages in rugose corals (or, hystero-ontogeny using the terminol- cation. Rather they could provide a favorable environmental back- ogy of rugosan students; see Smith and Ryder, 1926), brephic and ground that coincided with changes in composition of skeletog- sometimes early neanic stages of the protocorallite ontogeny can enous organic matrices (Cuif and Stolarski, 1999). Insights into be skipped (Oliver, 1968; Jull, 1973). However, it is reasonable such biologically driven changes of skeletal mineralogies are pro- to assume that in corals with rugose-style early ontogeny, hystero- vided by the stylasterids (Cnidaria, Hydrozoa), in which calcite ontogeny mirrors protocorallite ontogeny because in the hyster- vs. aragonite skeletal mineralogies may occur in different species ocorallite no stage of protocorallite ontogeny is skipped (compare and, interestingly, even in the ontogeny of one species (Cairns early hysterocorallites of Pachydendron in Cuif, 1975b, ®gs. 5a, and Macintyre, 1992). Calcite-aragonite changes in skeletal min- 7a, and early protocorallite of Pachythecalis in Cuif, 1975b, pl. eralogy also occur in the ontogeny of other invertebrates e.g., 13:2). In typical scleractinian corals, i.e., with simultaneous in- some bryozoans (see Sandberg, 1975) and ostreid bivalves (Me- sertion of six protosepta, hysterocorallites also mirror this stage dakovic et al., 1997). At least in mollusks, aragonite vs. calcite of septal development. polymorphism of shell layers was proven to be controlled by or- In the second analysis with Protoheterastraea as an outgroup, ganic macromolecules (Falini et al., 1996; Miyamoto et al., 1996; absence of one/two or axial septum is used as ancestral state. Choi and Kim, 2000). These data suggest that mineralogical dif- Character IV.Insertion of initial septa. The ancestral condi- ferences between the Rugosa and the Scleractinia may not be as tion of this character is considered to be two-by-two protoseptal important as thought by some authors as evidence against the insertion as found in solitary plerophyllines. hypothesis of the plerophylline → pachythecaliine transition (Ol- Simultaneous insertion of initial septa is assumed to be ances- iver, 1981; Fedorowski, 1997). In our analysis, calcitic mineralogy tral in the second analysis. is interpreted to be the plesiomorphic state, as it is found in pler- Character V.Metaseptal insertion. Metasepta are inserted se- ophyllines. rially in plerophylline rugosans and this character state is thus RONIEWICZ AND STOLARSKIÐTRIASSIC AMPHIASTRAEID CORALS 41

FIGURE 6ÐMitrodendron ogilviae Geyer, 1955. Lower Kimmeridgian (Upper Jurassic), SwieÎtokrzyskie Mountains (Holy Cross Mts.), Poland. ZPAL H.III/111. A series of transverse sections showing some stages of formation of a young corallite produced by Taschenknospung. Typical amphias- traeid prominence of an axial septum is shown in the right section. The new bud expands primarily into the cavity of the parental calice. Measurements (h) give vertical height above ®rst peel. considered plesiomorphic in the ®rst analysis. In the adult stage PAUP 4.0b (Swofford, 1998). Characters were coded as binary of Zardinophyllum, the metasepta appear to be inserted cyclically, variables (0,1) or as multistate characters (0, 1, 2), the 0 state but the actual insertion order is irregular and not resolvable into re¯ecting the presumed ancestral condition. Polymorphic states typical cyclic or serial (the problem is discussed in detail by Sto- were coded as (0/1) and missing values were indicated by ``?''. larski 1999). All multistate characters were treated unordered. In the second analysis with Protoheterastraea as outgroup cy- The ®rst analysis (plerophylline rugosans as an outgroup) re- clic metaseptal insertion is considered ancestral. sulted in 20 equally parsimonious trees generated in the exhaus- Character VI.Strong bilateral symmetry in the adult stage. tive search algorithm, each having 16 steps and a consistency In many corals, septa that appear ®rst in ontogeny retain their index (CI) of 0.8125. Successive weighting by maximum value prime character of being the largest also in the adult growth-stage. of rescaled consistency indicesÐa procedure for weighting char- In corals with one/two or axial septa in initial ontogeny, some or acters a posteriori according to their cladistic consistency (see all these septa (considered directive) are the largest, but also can Kitching et al., 1998)Ðreduces the number of trees to ®ve, with be disproportionately small or reduced. These septa confer a a consistency index CI ϭ 0.9732. In all ®ve trees the Quensted- strong bilateral symmetry on the corallite. The ancestral condition tiphylliinae (a new subfamily to enclose QuenstedtiphylliaÐsee is considered to be the presence of strong bilateral symmetry, as systematic paleontology section) is grouped with Amphiastraeinae found in plerophyllines. (the synapomorphy is Taschenknospung). The 50 percent majority In the second analysis a quasi-radial symmetry (absence of rule consensus tree of all ®ve trees (consensus tree formed from strong bilateral symmetry) as found in Protoheterastraea is con- those components that occur in at least 50 percent of ®ve trees) sidered ancestral. is illustrated as Figure 7.1. This tree is identical to one of ®ve Character VII.Morphology of the adult calice. In plerophyl- equally parsimonious trees and apomorphies from this tree are lines, septa are formed deeply in the calice, and this character added to this cladogram. state is assumed to be ancestral. The second analysis (Protoheterastraea as an outgroup; pler- Exsert septa are typical of Protoheterastraea (and the majority ophyllines removed) resulted in 5 equally parsimonious trees gen- of scleractinians) and this character state is considered ancestral erated in the exhaustive search algorithm, each having 13 steps and a consistency index (CI) of 0.8462. In all ®ve trees the Quen- in the second analysis. stedtiphylliinae is grouped with Amphiastraeinae (the synapo- Character VIII.Septal microstructure. The mid-septal zone of morphy is Taschenknospung). The 50 percent majority rule con- the plerophylline skeleton may consist of non-separated or sepa- sensus of all ®ve trees is illustrated as Figure 7.2. rated calci®cation centers, and we assume this polymorphic char- In both analyses, the clade (Amphiastraeinae and Quenstedti- acter state is ancestral in the ®rst analysis. Septa grow in an axial phylliinae) is supported by the same synapomorphy: Taschenk- direction. nospung type of asexual increase. Pachythecaliina (or, Hexanthi- In Protoheterastraea, calci®cation centers of the mid-septal niaria Montanaro-Gallitelli, 1975 considered as separate from the zone are well separated, and we assume that this is the ancestral anthozoan order Scleractinia), the larger clade that includes am- condition in the second analysis. phiastraeinae, quenstedtiphylliine and zardinophyllid genera, is Character IX.Wall. Presence of thick epitheca is the assumed supported by the synapomorphy: presence of the pachythecal ancestral state as found in plerophyllines (®rst analysis) and Pro- wall. toheterastraea (second analysis). Rationale for using microstructural characters in coral phy- Character X.Endotheca. Plerophyllines have endothecate and logenetic analysis.This is a ®rst and preliminary study of phy- non-endothecate coralla, and this polymorphic state is a plesiom- logenetic relationships of pachythecaliine corals; comprehensive orphic character in the ®rst analysis. Because Protoheterastraea analysis of all groups of Mesozoic corals considered to be related has endotheca, this character state is assumed ancestral in the to pachythecaliines and protoheterastreids will be published sep- second analysis. arately. There are several questions concerning usefulness of cla- Methods and results.Phylogenetic trees were generated using distic methodology for coral studies in general. Some workers 42 JOURNAL OF PALEONTOLOGY, V. 75, NO. 1, 2001 ces) using plerophylline ate: radial symmetry in adult stage). f, 1975b; Melnikova, 1986; Montanaro- Ï, 1987. Fifty-percent majority-rule consensus of ®ve trees obtained 2, Ïek and Ramovs as an outgroup (all characters have equal weight). This tree is identical to one of ®ve equally parsimonious trees and apomorphies from this tree are Protoheterastraea Fifty-percent majority-rule consensus tree of ®ve trees obtained in analysis (characters reweighted by maximum value of rescaled consistency indi 1, Ð 7 rugosans as an outgroup.This Numbers tree on is internode identical branches to indicate one the of percent ®ve of equally trees parsimonious that trees supports and each apomorphies node, from R.S. this is tree abbreviation are of added the to character this st cladogram. in analysis with added to this cladogram. Numbers on internode branches indicate the percent of trees that supports each node. Stratigraphic ranges of taxa after: Cui Gallitelli, 1975; Morycowa, 1971; Morycowa and Marcopoulou-Diacantoni, 1994; Roniewicz and Michalik, 1991; Turns IGURE F RONIEWICZ AND STOLARSKIÐTRIASSIC AMPHIASTRAEID CORALS 43

TABLE 1ÐCharacters and character states (1) and character matrix (2) used Discussion.More details about amphiastraeids can be found in the phylogenetic analyses of pachythecaliine families. The full set of characters was used in the ®rst analysis, which uses plerophylline rugosans in Beauvais (1974, 1976), Eliasova (1975, 1976, 1978), and as the outgroup. Bold characters and taxa are those that were removed from Koøodziej (1995). Among the amphiastraeids we distinguish two the second analysis, which uses Protoheterastraea as the outgroup. subfamilies: Amphiastraeinae Ogilvie, 1897 and Quenstedtiphyl- liinae subfam n. 1. Character Character states Subfamily AMPHIASTRAEINAE Ogilvie, 1897 I. Skeletal miner- 0. Calcitic; 1. Aragonitic  ology Diagnosis. Amphiastraeids with two-zonal endotheca. Coral- II. Budding 0. Absent (solitary); 1. Parietal-lateral; 2. Parietal- lites with bilateral symmetry in the adult stage. Taschenknospung Genera included.Amphiastraea EÂ tallon, 1859 (?Callovian, III. One/two or axi- 0. Present; 1. Absent Kimmeridgian±Aptian), Pleurostylina de Fromentel, 1856 (Ar- al initial sep- tum govien±Tithonian), Amphiaulastraea Geyer, 1955 (Tithonian±Al- IV. Insertion of ini- 0. Two-by-two; 1. Simulataneous bian), Aulastraea Ogilvie, 1897 (Tithonian), Mitrodendron Quen- tial septa stedt, 1881 (Oxfordian±Aptian), Pleurophyllia de Fromentel, V. Metaseptal in- 0. Serial; 1. Cyclic 1856 (Oxfordian±Tithonian), Hexapetalum EliasÏova, 1975 (Ti- sertion VI. Strong bilateral 0. Present; 1. Absent (radial symmetry) thonian), Hykeliphyllum EliasÏova, 1975 (Tithonian), Cheilosmilia symmetry in Koby, 1888 (Kimmeridgian±Tithonian). the adult cor- Occurrence.Jurassic (Oxfordian)±Cretaceous (Aptian); Eu- allite VII. Morphology of 0. Tube-like; 1. Septa exsert rope, Africa,? North America. the adult cal- ice Subfamily QUENSTEDTIPHYLLIINAE new subfamily VIII. Septal micro- 0. Non-separated c.c. (axialward growth); 1. Sepa- structure rated c.c. (axialward growth); 2. Separated c.c. Diagnosis.Amphiastraeids with one-zonal (tabular) endoth- (fan-like growth) eca. Corallites with quasi-radial symmetry in the adult stage. IX. Wall 0. Epitheca; 1. Pachytheca Genera included.Quenstedtiphyllia Melnikova, 1975 X. Endotheca 0. Absent; 1. One-zonal (tabular); 2. Two-zonal  (tabular and vesicular) Occurrence. Lower Carnian (Upper Triassic); Eurasia. 2. I. II. III. IV. V. VI. VII. VIII. IX. X. Genus QUENSTEDTIPHYLLIA Melnikova, 1975, emended herein Plerophyllina 0 0/1 0 0 0 0 0 0/1 0 0/1 Type species.Hexastraea fritschi Volz, 1896; ®gured in Volz, Zardinophyllum 0 0 0 0 0 0 0 0 1 0 1896, pl. 11, ®gs. 14±20, and herein, Figures. 1.1±1.3, 2, 3 (non Pachythecalis 1 0 0 0 1 1 0 0 1 0 Volzeia (Hexastraea) fritschi (Volz) in Cuif, 1975a, ®gs. 25, 26). Pachydendron 1 1 0 0 1 0 0 0 1 1  Pachysolenia 1 1 ? ? 1 1 0 ? 1 1 Species included. Quenstedtiphyllia fritschi (Volz, 1896); Quenstedtiphyllia 1 2 0 0 1 1 0 1 1 1 Quenstedtiphyllia mardjanaica Melnikova, 1975. Amphiastraeinae 1 2 0 0 1 0 0 1 1 2 Emended diagnosis.Phaceloid with quasi-radial symmetry in Protoheterastraea 1 0 1 (?) 1 1 1 1 2 0 1 the adult stage. Occurrence.Lower Carnian (Upper Triassic); Eurasia. Discussion.The generic name Hexastraea proposed by Volz argue that cladistics may not be successful, ®rstly, because ho- (1896) to contain two species, H. leonhardi and H. fritschi, was moplasy is very abundant and recurrent evolutionary trends were rejected by Wells (1937) as preoccupied, and was replaced by a recognized in many coral lineages (Webb, 1993, 1996), and sec- new name, Protoheterastraea. The lectotype specimen of Proto- ondly because coral evolution shows a reticulate pattern (Veron, heterastraea leonhardi (indicated by Roniewicz and Morycowa, 1995). Webb's (1993) arguments that convergent characters used 1993), ®gured many times (Volz, 1896, pl. 11, ®gs. 22, 23; Ron- in most coral studies are not compensated in phylogenetic analysis iewicz and Morycowa, 1993, ®g. 2.1±2.3; Stolarski, 1996a, ®g. by non-convergent characters, clearly hold for many morphologic 9a±c; Roniewicz and Stolarski, 1999, ®g. 7a, b; non Hexastraea characters (e.g., colony type and shape) which, almost exclusive- leonhardi Volz in Cuif, 1973, ®gs. 23±26), is well preserved. P. ly, have been used in phylogenetic analyses of corals. However, leonhardi is the monotypic species of the family Protoheterastei- many microstructural characters (e.g., pattern of distribution of daeae Cuif, 1977. The other species, Hexastraea fritschi, with a calci®cation centers) and those concerning skeletal ontogeny (e.g., new generic name Quenstedtiphyllia, was included by Melnikova early phases of skeleton formation) appear more stable during (1975) in the family Amphiastraeidae Ogilvie, 1897. This species coral evolution (but see discussion in ``Biomineralization and its was illustrated only once, in the original description by Volz implication'' chapter and Stolarski, 2000), and we believe that (1896, pl. 11, ®gs. 14±20). their use may help to re®ne application of the cladistic approach Among the pachythecaliines, Quenstedtiphyllia resembles to corals. Veron's (1995) concept of reticulate evolution, though Pachysolenia Cuif, 1975. Features in common with Pachysolenia of potentially great impact on coral evolutionary studies, still are: phaceloid growth form, radial symmetry, and tabular endoth- needs to be con®rmed by empirical studies like those by Kenyon eca. Differences between them lie in: (1) the microstructure of (1997) on Acropora. the pachytheca, which in Quenstedtiphyllia is built of modules that are centered on the wall (Fig. 2.3), not along subhorizontal SYSTEMATIC PALEONTOLOGY axes (Fig. 2.4), as in Pachysolenia (and zardinophyllids), (2) the Suborder PACHYTHECALIINA EliasÏova, 1976 emended herein mode of budding, which in Quenstedtiphyllia is a typical Tas- Emended diagnosis.Corals with pachytheca. chenknospung and different from the lateral budding observed in Pachysolenia, and possibly (3) in the microstructure of the septa: Family AMPHIASTRAEIDAE Ogilvie, 1897 emended herein the non-separated calci®cation centers of the mid-septal zone in Emended diagnosis.Pachythecaliines with the Taschenknos- Pachysolenia (see remarks on septal microstructure, above in the pung type of asexual increase. Separated septal calci®cation cen- text) are different from the separated calci®cation centers in am- ters. phiastraeids. 44 JOURNAL OF PALEONTOLOGY, V. 75, NO. 1, 2001

CONCLUSIONS Stylophyllidae. Bulletin du MuseÂum National d'Histoire Naturelle, Sci- 1) The Early Carnian (Upper Triassic) phaceloid Quenstedti- ences de la Terre, 17 (for 1972):211±291. CUIF, J. P. 1975a. Recherches sur les MadreÂporaires du Trias. II. Astraeo- phyllia fritschi (Volz, 1896) reproduced asexually by Taschenk- ida. ReÂvision des genres Montlivaltia et Thecosmilia. Etude de qu- nospung (pocket-budding), a character considered an autapomor- elques types structuraux du Trias de Turquie. Bulletin du MuseÂum Na- phy for Amphiastraeidae. tional d'Histoire Naturelle, Sciences de la Terre, 40 (for 1974):293± 2) New polyps created by Taschenknospung became separated 400. from the parental polyp after digestive (mesenteria) and defense CUIF, J. P. 1975b. CaracteÁres morphologiques, microstructuraux et sys- (tentacular crown) systems were well developed. They thus were teÂmatiques des Pachythecalidae nouvelle famille de MadreÂporaires tria- possibly better adapted to feeding and less vulnerable to predators siques. GeÂobios, 8:157±180. than polyps created by lateral budding. CUIF, J. P. 1976. Recherches sur les MadreÂporaires du Trias. Formes 3) Changes in biomineralization styles that occur in blastogeny ceÂrio-meÂandroides et thamnasteÂrioides du Trias des Alpes et du Taurus sud-anatolien. Bulletin du MuseÂum National d'Histoire Naturelle, Sci- of Q. fritschi (modular vs. trabecular style of biocalci®cation) ences de la Terre, 53:65±195. provide important clues about possible constraints on the micro- CUIF, J. P. 1977. Arguments pour une relation phyleÂtique entre les Mad- structural evolution of the scleractinian skeleton. This casts light reÂporaires paleÂozoõÈques et ceux du Trias. Implications systeÂmatiques on possible phylogenetic relationships between corals with dif- de l'analyse microstructurale des MadreÂporaires triasiques. MeÂmoires ferent types of skeletal microstructures (e.g., zardinophyllids and de la SocieÂteÂGeÂologique de France, n.s., 56 (129):1±54. protoheterastraeids), which have thus far been considered unre- CUIF,J.P.,AND Y. D AUPHIN. 1998. Microstructural and physico-chemical lated. characterization of `centers of calci®cation' in septa of some Recent 4) Quenstedtiphyllia fritschi shares several plesiomorphies with scleractinian corals. PalaÈontologische Zeitschrift, 72:257±270. Triassic Zardinophyllidae: e.g., very thick, modular epithecal wall CUIF,J.P.,AND J. STOLARSKI. 1999. Origin and paleobiology of wall- th (pachytheca) and strongly bilateral early blastogenetic stages with based corals. Abstracts of the 8 International Symposium on Fossil the earliest corallite having one axial initial septum. One-zonal- Cnidaria and Porifera, Sendai, 95. DENG ZHANQIU, AND KONG LEI. 1984. Middle Triassic corals and spong- endotheca (vs. two-zonal in Amphiastraeinae) is a distinguishing es from Southern Guizhou and Eastern Yunnan. Acta Palaeontologica character of the new amphiastraeid subfamily Quenstedtiphylli- Sinica, 23:489±504. (In Chinese, with English summary) inae. Phylogenetic analysis of zardinophyllids and amphiastraeids ELIASÏOVA, H. 1975. Sous-ordre Amphiastraeina Alloiteau, 1952 (Hexa- supports the grouping of Zardinophyllidae with Amphiastraeidae corallia) des calcaires de SÏtramberk (Tithonien, TcheÂcoslovaquie). CÏ a- in the clade Pachythecaliina (synapomorphy: presence of pachy- sopis pro mineralogii a geologii, 20:1±23. theca). ELIASÏOVA, H. 1976. Les coraux de l'ordre Hexanthiniaria Montanaro- Gallitelli, 1975, Zoantharia de Blainville, 1830 dans les calcaires de ACKNOWLEDGMENTS SÏtramberk (Tithonien, TcheÂcoslovaquie). VeÏstnik U strÏedniho uÂstavu The work has been accomplished in the Institute of Paleobi- geologickeÂho, 51:357±366. ELIASÏOVA, H. 1978. La rede®nition de l'ordre Hexantiniaria Montanaro- ology, Warsaw with the support of the Committee for Scienti®c  Research (KBN) grant 6 P04D 037 14 to E. Roniewicz and J. Gallitelli, 1975 (Zoantharia). VeÏstnik UstrÏedniho uÂstavu geologickeÂho, 53:89±101. Stolarski, and in the National Museum of Natural History, Wash- E TALLON, A. 1859. E tudes paleÂontologiques sur le Haut-Jura. RayonneÂes ington, D.C. during a postdoctoral fellowship to J. Stolarski grant- du Corallien. MeÂmoires de la SocieÂte d'E mulation du Departement du ed by Smithsonian Institution. We sincerely thank G.D. Stanley, Doubs, seÂr. 3, 3:401±550. Jr. who provided us with the specimen of Pachysolenia cylindrica. EZAKI, Y. 1997. The Permian coral Numidiaphyllum: new insights into We appreciate very much remarks provided by Brian Rosen (Nat- anthozoan phylogeny and Triassic scleractinian origins. Palaeontology, ural History Museum, London), and Bernard Lathuiliere (Univ- 40:1±14. ersiteÁ de Nancy) who reviewed this paper. We owe thanks for FEDOROWSKI, J. 1997. Rugosa and ScleractiniaÐa commentary on some technical support to Z. Strak, A. Hoøda, G. DziewinÂska, and M. methods of phylogenetic reconstructions. Acta Palaeontologica Polon- DziewinÂski (all from the Institute of Paleobiology, Warsaw). ica, 42:446±456. FROMENTEL,E.DE 1856. Note sur les polypiers fossiles de l'eÂtage por- REFERENCES tlandien de la Haute-SaoÃne. Bulletin de la SocieÂteÂgeÂologique de France, seÂr. 2, 13:851±864. BEAUVAIS, L. 1974. Studies upon the characters of the suborder Am- GEYER, O. 1955. BetraÈge zur Korallenfauna des Stramberger Titon. Pa- phiastraeida Alloiteau, p. 238±248. In B. S. Sokolov (ed.), Ancient Cnidaria 1. Nauka, Novosibirsk. laÈontologische Zeitschrift, 29:177±216. HILL, D. 1981. Rugosa and Tabulata, p. F5-F743. In C. Teichert (ed.), BEAUVAIS, L. 1976. Etude morphologique, taxonomique et phylogeÂneÂ- tique du sous-ordre Amphiastraeida Alloiteau. MeÂmoires de la SocieÂte Treatise on Invertebrate Paleontology, Pt. F (Coelenterata), Supplement GeÂologique de France, nouvelle seÂr., 50 (126):5±42. I. Geological Society of America and University of Kansas Press, Lawrence. CAIRNS, S. D. 1995. The marine fauna of New Zealand: Scleractinia (Cnidaria: Anthozoa). New Zealand Oceanographic Institute Memoir, ILINA, T. G. 1984. Historical development of Anthozoa. Suborder Poly- 103:1±210. coeliina. Trudy Paleontologiceskogo Instituta, 198:1±183. (In Russian) CAIRNS,S.D.AND I. G. MACINTYRE. 1992. Phylogenetic implications JELL, J. S. 1981. Skeletogenesis of newly settled planulae of the herma- of calcium carbonate mineralogy in the Stylasteridae (Cnidaria: Hy- typic coral Porites lutea. Acta Palaeontologica Polonica, 25 (for 1980): drozoa). Palaios, 7:96±107. 311±320. CHOI, C. S., AND Y. W. K IM. 2000. A study of the correlation between JULL, R. K. 1973. Ontogeny and hystero-ontogeny in the Middle Devo- organic matrices and nanocomposite materials in oyster shell forma- nian rugose coral Hexagonaria anna (Whit®eld), p. 59±68. In R. S. tion. Biomaterials, 21:213±222. Boardman, A. H. Cheetham and W. A. Oliver, Jr. (eds.), Animal Col- COATES,A.G.,AND J. B. C. JACKSON. 1985. Morphological themes in onies. Dowden, Hutchinson and Ross, Stroudsburg. the evolution of clonal and aclonal marine invertebrates, p. 67±106. In KENYON, J. C. 1997. Models of reticulate evolution in the coral genus J. B. C. Jackson, L. W. Buss, and R. E. Cook (eds.), Biology and Acropora based on chromosome numbers: parallels with plants. Evo- Evolution of Clonal Organisms. Yale University Press, New Haven. lution, 51:756±767. CONSTANTZ, B. R., AND A. MEIKE. 1990. Calcite centers of calci®cation KITCHING, I. J., P. L. FOREY,C.J.HUMPHRIES, AND D. M. WILLIAMS. in Mussa angulosa (Scleractinia), p. 201±207. In R. E. Crick (ed.), 1998. Cladistics: The Theory and Practice of Parsimony Analysis (sec- Origin, Evolution, and Modern Aspects of Biomineralization in Plants ond edition). Oxford University Press, Oxford, 223 p. and Animals. Plenum Press, New York. KOBY, F. 1888. Monographie des polypiers jurassiques de la Suisse. MeÂ- CUIF, J. P. 1973. Recherches sur les MadreÂporaires du Trias. I. Famille moires de la SocieÂte PaleÂontologique Suisse, 25:401±456. RONIEWICZ AND STOLARSKIÐTRIASSIC AMPHIASTRAEID CORALS 45

KOèODZIEJ, B. 1995. Microstructure and taxonomy of Amphiastraeina in the light of microstructural data. Courier Forschungsinstitut Senck- (Scleractinia). Annales Societatis Geologorum Poloniae, 65:1±17. enberg, 164:233±240. MEDAKOVIC, D., S. POPOVIC,B.GRZETA,M.PLAZONIC, AND M. HRS- RONIEWICZ, E., AND J. STOLARSKI. 1999. Evolutionary trends in the ep- BRENKO. 1997. X-ray diffraction study of calci®cation processes in ithecate scleractinian corals. Acta Palaeontologica Polonica, 44:131± embryos and larvae of the brooding oyster Ostrea edulis. Marine Bi- 166. ology, 129(4):615±623. ROSEN, B. R. 1986. Modular growth and form of corals: a matter of MELNIKOVA, G. K. 1975. Late Triassic Scleractinians of the South-West- metamers?. Philosophical Transactions of the Royal Society of London, ern Pamirs. p. 1±234, Donish, Dushanbe. (In Russian) B 313:115±142. MELNIKOVA, G. K. 1986. New data on systematics and phylogeny of the SANDBERG, P. A. 1975. Bryozoan diagenesis: bearing on the nature of pachythecaliids (Scleractinia). In B. S. Sokolov (ed.), Fanerozoiskie the original skeleton of rugose corals. Journal of Paleontology, 49:587± rify i korally SSSR. Nauka, Moskva. (In Russian) 606. CHINDEWOLF MELNIKOVA,G.K.,AND E. RONIEWICZ. 1976. Contribution to the sys- S , O. 1942. Zur Kenntnis der Polycoelien und Plerophyllen. tematics and phylogeny of Amphiastraeina (Scleractinia). Acta Pa- Eine Studie uÈber den Bau der ``Tetrakorallen'' und ihre Beziehungen laeontologica Polonica, 21:97±114. zu den Madreporarien. Abhandlungen des Reichsamts fuÈr Bodenfor- schung, Neue Folge, 204:1±324. MIYAMOTO, H., T. MIYASHITA,M.OKUSHIMA,S.NAKANO,T.MORITA, SMITH, S., AND T. A. RYDER. 1926. The genus Corwenia, gen. nov. An- AND A. MATSUSHIRO. 1996. A carbonic anhydrase from the nacreous nals and Magazine of Natural History, (9) 17:149±159. layer in oyster pearls. Proceedings of the National Academy of Sci- SORAUF, J., AND CUIF, J. P. 1999. Diagenesis in Scleractinia. Abstracts ences of the United States of America, 93:9657±9660. of the 8th International Symposium on Fossil Cnidaria and Porifera, MONTANARO-GALLITELLI, E. 1975. Hexanthiniaria a new ordo of Zoan- Sendai, 45. tharia (Anthozoa, Coelenterata). Bollettino della SocietaÁ Paleontologica STANLEY,S.M.,AND L. A. HARDIE. 1998. Secular oscillations in the Italiana, 14(1, preprint):21±25. carbonate mineralogy of reef-building and sediment-producing organ- MORYCOWA, E. 1971. Hexacorallia et Octocorallia du CreÂtace infeÂrieur isms driven by tectonically forced shifts in seawater chemistry. Pa- de RaraÏu (Carpathes orientales roumaines). Acta Palaeontologica Po- laeogeography, Palaeoclimatology, Palaeoecology, 144:3±19. lonica, 16:1±149. STOLARSKI, J. 1996a. GardineriaÐscleractinian ``living fossil''. Acta Pa- MORYCOWA, E., AND A. MARCOPOULOU-DIACANTONI. 1994. Cretaceous laeontologica Polonica, 41:339±367. scleractinian corals from the Parnassos area (Central Greece) (prelim- STOLARSKI, J. 1996b. Paleogene corals from Seymour Island, Antarctic inary note). Bulletin of the Geological Society of Greece, 30:249±273. Peninsula. In A. GazÂdzicki (ed.), Palaeontological Results of the Polish OGILVIE, M. 1896. Microscopic and systematic study of madreporarian Antarctic Expeditions. Pt. II. Palaeontologia Polonica, 55:51±63. types of corals. Philosophical Transactions of the Royal Society, B 187: STOLARSKI, J. 1999. Early ontogeny of the skeleton of Recent and fossil 83±345. Scleractinia and its phylogenetic signi®cance. Abstracts of the 8th In- OGILVIE, M. 1897. Die Korallen der Stramberger Schichten. Palaeonto- ternational Symposium on Fossil Cnidaria and Porifera, Sendai, 37. graphica (Suppl. 2), 7:76±284. STOLARSKI, J. 2000. Origin and phylogeny of Guyniidae (Scleractinia) in OLIVER, W. A., JR. 1968. Some aspects of colony development in corals. the light of microstructural data. Lethaia, 33:13±38. Journal of Paleontology, 42:16±34. SWOFFORD, D. L. 1998. PAUP, Phylogenetic Analysis Using Parsimony, OLIVER, W. A., JR. 1981. On the relationship between Rugosa and Scler- version 4. 0b2 for Macintosh. Sinauer Associates, Inc. Ï Ï actinia (Summary). Acta Palaeontologica Polonica, 25(for 1980):339± TURNSEK, D., AND RAMOVS, A. 1987. Upper Triassic (Norian-Rhaetian) 367. reef buildups in the northern Julian Alps (NW Yugoslavia). Rozprave Slovenska Akademija Znanosti in Umetnosti 28(2):27±67. QUENSTEDT, F. A. 1881. Petrefactenkunde Deutschlands, 6: Korallen. VERON, J. E. N. 1995. Corals in Space and Time: Biogeography and Fues Verlag, Lepzig, 1093 p. Evolution of the Scleractinia. Cornell University Press, Ithaca, N.Y., RAILSBACK, L. B., AND T. F. A DERSON. 1987. Control of Triassic seawater 321 p. chemistry and temperature on the evolution of post-Paleozoic arago- VOLZ, W. 1896. Die Korallenfauna der Trias. II. Die Korallen der Schich- nite-secreting faunas. Geology, 15:1002±1005. ten von St. Cassian in SuÈd Tirol. Palaeontographica, 443:1±124. RINKEVICH, B., AND Y. L OYA. 1979. The reproduction of the Red Sea WEBB, G. E. 1993. Phylogeny reconstruction: problems posed by Paleo- coral Stylophora pistillata. I. Gonads and planulae. Marine Ecology zoic corals. Courier Forschungsinstitut Senckenberg, 164:71±74. Progress Series, 1:133±144. WEBB, G. E. 1996. Morphological variation and homoplasy: the challenge RONIEWICZ, E. 1966. Les MadreÂporaires du Jurassique supeÂrieur de la of Paleozoic coral systematics. Paleontological Society Papers, 1:135± bordure des Monts de Sainte-Croix, Pologne. Acta Palaeontologica Po- 157. lonica, 11:157±264. WELLS, J. W. 1937. New genera of Mesozoic and Cenozoic corals. Jour- RONIEWICZ, E. 1996. The key role of skeletal microstructure in recog- nal of Paleontology, 11:73±77. nizing high-rank scleractinian taxa in the stratigraphical record. Pale- WENDT, J. 1990. The ®rst aragonitic rugose coral. Journal of Paleontol- ontological Society Papers, 1:187±206. ogy, 64:335±340. RONIEWICZ, E., AND J. MICHALIK. 1991. Zardinophyllum (Scleractinia) YAMASHIRO, H., AND K. YAMAZATO. 1987. Studies on the detachment from the Upper Triassic of the Central Western Carpathians (Czecho- of the disc of the mushroom coral Fungia fungites with special refer- Slovakia). Geologica Carpathica, 42:361±363. ence to hard structural changes. Galaxea, 6:163±175. RONIEWICZ, E., AND E. MORYCOWA. 1993. Evolution of the Scleractinia ACCEPTED 21 AUGUST 2000